How are you monitoring your infrastructure?

HalfEatenPie

@tjn said: (I don't usually like tagging people but might resort to it 😅)

I'd recommend against tagging people unless they've already posted in this thread.

For self-hosted distributed monitoring I'd also suggest nagios, centreon (usually a frontend to nagios), icinga (historically built on top of nagios but 2.0 was built from the ground up).

But distributed monitoring is... In my opinion, not worth it as much. It adds more operational overhead and can tell you a big more on which connections have problems but actual actions around that information is limited (nothing beyond raising a ticket and even then... unless you're a bigger host what do you expect them to do?). Imho, unless you're building out your own hard-out network infrastructure as a host or even as a hobbiest it doesn't really matter that much.

bulbasaur

I use Grafana for visualization, Prometheus for metrics collection and Loki for log ingestion.

I also monitor node uptimes through Freshping.

Mumbly

@tjn said:

@Mumbly said:
I used Newrelic for few months.

I used to use it too a few years back. It was awesome then got expensive. Haven't tried it since.

They advertise it as "free forever" now. It's really complex and powerful tool if they just wouldn't nag me with all sort of "advice" mails all the time even after I opted out from receiving them.
I could simply filter them or use some half-abandoned mail address but this would beat the purpse to have so powerful monitoring tool installed at all as there may be some legit alert notifications via email.

TimboJones

@Mumbly said:

@tjn said:

@Mumbly said:
I used Newrelic for few months.

I used to use it too a few years back. It was awesome then got expensive. Haven't tried it since.

They advertise it as "free forever" now. It's really complex and powerful tool if they just wouldn't nag me with all sort of "advice" mails all the time even after I opted out from receiving them.
I could simply filter them or use some half-abandoned mail address but this would beat the purpse to have so powerful monitoring tool installed at all as there may be some legit alert notifications via email.

Opt out worked for me. Not sure why you're not getting removed. No spam hassles after the initial sign-up flurry, par for the course, of course.

Mumbly

@TimboJones said:

@Mumbly said:

@tjn said:

@Mumbly said:
I used Newrelic for few months.

I used to use it too a few years back. It was awesome then got expensive. Haven't tried it since.

They advertise it as "free forever" now. It's really complex and powerful tool if they just wouldn't nag me with all sort of "advice" mails all the time even after I opted out from receiving them.
I could simply filter them or use some half-abandoned mail address but this would beat the purpse to have so powerful monitoring tool installed at all as there may be some legit alert notifications via email.

Opt out worked for me. Not sure why you're not getting removed. No spam hassles after the initial sign-up flurry, par for the course, of course.

At some point they started with "advice newsletters" numbered as 1/5 (first out of five) or something like that every few days one so I went through all mailing settings again just to be sure. Everything still opted out, so I thought "ok, it is what it is, I will survive those five mails" and then after the 5th mail arrived just new 1/5 round started. I was so fed up of them that I just cancelled

nessa

Does anyone still use Nagios lol

HalfEatenPie

@nessa said:
Does anyone still use Nagios lol

There are still a fair amount of people and organizations who use nagios.

TimboJones

@Mumbly said:

@TimboJones said:

@Mumbly said:

@tjn said:

@Mumbly said:
I used Newrelic for few months.

I used to use it too a few years back. It was awesome then got expensive. Haven't tried it since.

They advertise it as "free forever" now. It's really complex and powerful tool if they just wouldn't nag me with all sort of "advice" mails all the time even after I opted out from receiving them.
I could simply filter them or use some half-abandoned mail address but this would beat the purpse to have so powerful monitoring tool installed at all as there may be some legit alert notifications via email.

Opt out worked for me. Not sure why you're not getting removed. No spam hassles after the initial sign-up flurry, par for the course, of course.

At some point they started with "advice newsletters" numbered as 1/5 (first out of five) or something like that every few days one so I went through all mailing settings again just to be sure. Everything still opted out, so I thought "ok, it is what it is, I will survive those five mails" and then after the 5th mail arrived just new 1/5 round started. I was so fed up of them that I just cancelled

A common problem with unsubscribes is that they're often on ad blocking DNS lists (email tracker domains). I'd screenshot and contact them if I thought I missed a removal properly twice.

JasperNL

How I monitor my services?
If a service is down, people start messaging me.

rcy026

@nessa said:
Does anyone still use Nagios lol

A lot of people do.
Nagios, Naemon and Icinga are still major players when you get to people that actually learn a product, not just install it.

nessa

@rcy026 said:

@nessa said:
Does anyone still use Nagios lol

A lot of people do.
Nagios, Naemon and Icinga are still major players when you get to people that actually learn a product, not just install it.

well we use Nagios for our internal monitoring just looked like maybe others have moved on the fancier things.

v3ng

@rcy026 said:

@v3ng said:

@rcy026 said:

@v3ng said:
I'm primarily using CheckMK for uptime + resource monitoring.
If a higher granularity is needed, I'm combing it with Netdata.

Unfortunately, CheckMK lacks a public dashboard, which is why I'm still stuck with Uptimerobot.

I've actually never tried it, but you can use CheckMK as a direct datasource in Grafana. You should be able to easily build a dashboard exactly the way you want it with the data you already have.
https://docs.checkmk.com/latest/en/grafana.html

I was not aware of that feature, thank you!

If you check it out, please let me know what you find. Trying out checkmk and grafana integration has been on my todo-list for a long time since I use both products, just haven't got around to it yet.

I'm currently playing around with it, however I have yet to figure out how to include multiple hosts per dashboard.

Otherwise, it's really good!

donko

HetrixTools, easy and free

szarka

@nessa said:
I think we've generally been happy with Nodeping from a monitoring perspective but I dont really like the status page

Same here. Haven't made any status pages public, but Nodeping is a great value for the price. I use it to monitor/alert uptime, SMTP, etc. I especially like that it has SSL cert monitoring, now that I'm using Let's Encrypt. They even have an API if you want to integrate it into your scripts, though I'm small enough that I haven't bothered yet.

rcy026

@v3ng said:

I'm currently playing around with it, however I have yet to figure out how to include multiple hosts per dashboard.

That can be accomplished in several ways. You could simply put values from multiple hosts in the same database and just use hostname as a separate value, or you put hostname as a selectable variable for the entire dashboard. Or several other ways, depending a bit on what your datasource supports.
Just pm me if you need some help, playing with grafana is always fun.

rcy026

@nessa said:

well we use Nagios for our internal monitoring just looked like maybe others have moved on the fancier things.

Nagios might not be fancy, but when it comes to raw monitoring it's quite hard to beat.
Use the data from Nagios to feed a Grafana or Kibana frontend, and its about as fancy as it gets.

analog

I use Zabbix inhouse then uptimerobot to monitor from external

c1vhosting

@tjn said:
Curious about what everyone is using these days to monitor their servers and infrastructure?

Do you outsource to the likes of Nixstats, Site24x7, or HetrixTools?
Or "self-monitor" with something like Zabbix, Checkmk, Prometheus+Influx+Grafana, PRTG, Observium, etc...

If you do it in-house, are you monitoring from multiple locations? If so, how?

PS - I'm not asking about website uptime monitoring like pingdom and uptimerobot.

Our NOC monitors our core infrastructure. Our servers are automatically checked by multiple local and remote "uptime checker" and if there's a issue then it's checked manually.

v3ng

@rcy026 said:

@v3ng said:

I'm currently playing around with it, however I have yet to figure out how to include multiple hosts per dashboard.

That can be accomplished in several ways. You could simply put values from multiple hosts in the same database and just use hostname as a separate value, or you put hostname as a selectable variable for the entire dashboard. Or several other ways, depending a bit on what your datasource supports.
Just pm me if you need some help, playing with grafana is always fun.

Yep but when using CheckMK as a datasource you only seem to be able to select a single host per graph.

Will dig deeper tomorrow, maybe I missed some option

rcy026

@v3ng said:
Yep but when using CheckMK as a datasource you only seem to be able to select a single host per graph.

Will dig deeper tomorrow, maybe I missed some option

I might have to setup a CheckMK and check it out. As I said, it's been on my todo-list for a while anyway.